Distribution of data in cellular electrophysiology: Is it always normal?

ABSTRACT

The distribution of data presented in many electrophysiological studies is presumed to be normal without any convincing evidence. To test this presumption, the cell membrane capacitance and magnitude of inward rectifier potassium currents were recorded by the whole-cell patch clamp technique in rat atrial myocytes. Statistical analysis of the data showed that these variables were not distributed normally. Instead, a positively skewed distribution appeared to be a better approximation of the real data distribution. Consequently, the arithmetic mean, used inappropriately in such data, may substantially overestimate the true mean value characterizing the central tendency of the data. Moreover, a large standard deviation describing the variance of positively skewed data allowed 95% confidence interval to include unrealistic negative values. We therefore conclude that the normality of the electrophysiological data should be tested in every experiment and, if rejected, the positively skewed data should be more accurately characterized by the median and interpercentile range or, if justified (namely in the case of log-normal and gamma data distribution), by the geometric mean and the geometric standard deviation.